Percutaneous Electrical Nerve Field Stimulation Improves Weight and Other Anthropometric Outcomes in Children and Adolescents With Functional Dyspepsia Who Restrict Their Oral Intake

Abstract

Background:

Functional dyspepsia (FD), a disorder of the gut-brain interaction (DGBI), can result in restricted eating habits with associated weight loss. Percutaneous Electrical Nerve Field Stimulation (PENFS), a minimally invasive non-pharmacologic approach, can improve abdominal pain and psychological functioning in children with DGBI. We aimed to analyze changes in anthropometric measures, nutritional intake, and clinical outcomes with PENFS.

Methods:

We retrospectively reviewed charts of patients with FD (9–28 years) who restricted oral intake and underwent PENFS. We analyzed validated questionnaires used during treatment. Anthropometric data were also assessed during and after 3 months of treatment. Results were analyzed through linear mixed modeling as least square means.

Key Results:

Ninety patients reported restricted oral intake, 42% had baseline weight loss and 13% requiring enteral support. Oral intake (p =0.001), weight (p =0.02) and BMI (p =0.002) significantly increased after PENFS. Abdominal pain, nausea, functional disability, depression, and somatization also improved after PENFS. Increased weight Z scores were associated with lower nausea, pain catastrophizing, and depression.

Conclusions and Inferences:

PENFS improved oral intake, weight, and other anthropometric measures in addition to GI and psychological measures in patients with FD. Changes in anthropometrics were sustained at 3-month follow up. PENFS may be an effective treatment for patients with FD, including children, and restrictive eating to optimize nutritional outcomes. Future studies should include prospective clinical trials and investigate the mechanism of action of PENFS on eating habits.

1 |. Introduction

Functional dyspepsia (FD) is one of the most common disorders of gut-brain interaction (DGBI) within the pediatric population, affecting up to 4.5%–10% of children [1]. It is multifactorial in nature and has been related to visceral hypersensitivity, delayed gastric emptying, and vagal insufficiency. It is also influenced by environmental, genetic, and psychological factors [2]. Diagnosis of FD relies on Rome IV criteria that include specific symptoms of epigastric pain, fullness, and early satiety that can additionally be subcategorized into different subtypes (i.e., post-prandial distress syndrome, epigastric pain syndrome) [3].

Patients often restrict oral intake due to symptoms associated with FD, resulting in weight loss and nutritional deficiencies [4]. Furthermore, patients with FD that restrict oral intake have been associated with more severe symptoms and avoidant/restrictive food intake disorder (ARFID) [5]. Per DSM-5 criteria, ARFID is defined as an eating or feeding disturbance with a failure to meet nutritional needs that is not explained by lack of food, comorbid eating disorder, or other medical diagnosis [5]. Refractory or difficult-to-treat FD with possible ARFID can often lead to enteral nutritional needs, which can lead to a more prolonged and difficult treatment course [6].

As a result, finding effective treatments for FD with restricted oral intake is critical as it can have an impact on a patient’s nutritional health and quality of life. Although patients may demonstrate some benefit to medications and psychological care, these are typically sub-optimal alone for managing FD with restricted oral intake [7].

Auricular percutaneous electrical nerve field stimulation (PENFS) is a minimally invasive non-pharmacological approach involving modulation of central pain pathways through the external ear [8]. PENFS can significantly improve DGBI symptoms, including abdominal pain and nausea, as well as psychological functioning [9]. However, treatment outcomes specifically in pediatric populations with FD and ARFID have not yet been studied.

We aimed to examine changes in clinical and anthropometric outcomes of pediatric patients with FD and ARFID undergoing PENFS. We hypothesized that PENFS would improve clinical and anthropometric outcomes and nutritional intake in those with FD and ARFID. Additionally, patients on enteral nutrition would be able to improve their oral intake and decrease tube feeding dependence.

2 |. Methods and Materials

After obtaining institutional review board approval, the electronic medical records of patients ages 9–28years, who met the Rome IV criteria for FD and had been treated with at least 4weeks of PENFS between January 1, 2017 and December 31, 2023, were evaluated for restricted oral intake. Patients were considered to have restricted oral intake if they demonstrated or endorsed the following themes: picky eating, decreased appetite, having to push themselves to eat enough, avoiding certain foods or eating altogether due to GI symptoms (i.e., abdominal pain, nausea, vomiting, bloating) or sensory characteristics. The criteria were selected primarily using the Nine Item Avoidant/Restrictive Food Intake Disorder (ARFID) Screen (NIAS) that is a brief, reliable tool for capturing ARFID or ARFID-like eating behaviors [10]. Patients diagnosed with eating disorders (anorexia nervosa, bulimia nervosa, binge-eating disorder or other specified feeding and eating disorder), organic GI and systemic disorders, developmental delays or contraindications to PENFS [allergies to components of the device (metal lead, adhesive), hematological disorders causing excessive bleeding/bruising, and cardiac pacemakers] were excluded from the study.

Demographic data and medical history were obtained from the medical records. Medication use during PENFS was reviewed, specifically medications with known weight loss (amphetamine-dextroamphetamine, atomoxetine, dexmethylphenidate, lisdexafetamine, and methylphenidate) and weight gain (aripiprazole, cyproheptadine, topiramate, mirtazapine, olanzapine, and quetiapine).

Anthropometric data was collected at initial and subsequent weekly visits, as well as 3 months after PENFS treatment. Measures included weight, weight-for-age Z-scores (following World Health Organization), Body Mass Index (BMI), and BMI Z-scores. Additionally, the form of intake (oral, enteral, and parenteral) was examined before and after PENFS. Available dietitian notes were reviewed to record changes in the amount and type of food during treatment.

Symptoms and psychological outcomes were evaluated using validated questionnaires that were collected as a part of clinical care during visits. These included:

2.1 |. Abdominal Pain Index (API)

The API is validated for children to determine the frequency, severity, intensity and duration of abdominal pain over 2 weeks using Likert scales [11]. We used a modified version to reflect composite scores over 1 week.

2.2 |. Nausea Severity Scale (NSS)

The NSS is a reliable tool to determine the number of days with nausea, daily episodes and characteristics including duration and intensity over 2weeks [12]. This measure normally assesses nausea severity over 2weeks. We used a modified version with a duration of over 1week.

2.3 |. Pain Catastrophizing (PCS-C)

The PCS-C is a questionnaire for children and adolescents designed to measure how strongly they agree with maladaptive beliefs about pain and long-term outcomes [13]. We used the questionnaire for over 1 week.

2.4 |. Functional Disability Inventory (FDI)

The FDI is a validated instrument to clinically measure daily physical and psychosocial functioning due to impaired physical health [14]. While the measure normally assesses perceptions of activity limitations over 2weeks, we used a modified version for a duration of 1 week.

2.5 |. Pediatric Insomnia Severity Index (PISI)

PISI is a validated tool to measure the burden of insomnia in children and adolescents [15]. We used this for a duration of over 1 week.

2.6 |. Pain Coping Questionnaire (PCQ)

PCQ is a reliable tool used to understand how a child copes with pain of extended duration [16]. We used this for a duration of over 1week.

2.7 |. Anxiety (PROMIS Measure)

Changes in symptoms of anxiety were assessed through the short Patient Reported Outcomes Measurement Systems (PROMIS) [17]. We used this for a duration of over 1week.

2.8 |. Depression (PHQ-9)

Depressive symptoms were assessed through the Patent Health Questionnaire—9 (PHQ-9) [18]. We used this for a duration of over 1 week.

2.9 |. Children’s Somatization Inventory (CSI)

CSI is a reliable tool used to examine somatization of general and gastrointestinal-specific pain symptoms a child with chronic abdominal pain can experience [19]. We used this for a duration of over 1 week.

Changes in anthropometrics and outcomes were measured immediately after and at 3 months post-completion of 4weeks of treatment.

2.10 |. Statistical Analysis

Descriptive statistics were presented as mean (with standard deviation) or median with inter-quartile range for continuous and as frequency count(%) for categorical data. Two anthropometric measures, namely weight and BMI, were log-transformed and results were presented after back-transformation. Linear mixed modeling was used to examine questionnaire scores longitudinally by placement visits, with visits modeled as class variable and fixed effect and subjects as random effect to account for repeated measures. Results were reported as least square means (LSMeans) and 95% confidence interval (95% CI). p value for change in questionnaire scores with subsequent visits was examined by modeling visits as a continuous variable. Similarly, anthropometric measures were also examined longitudinally using linear mixed models and results reported as LSMeans (95% CI). For anthropometric measures, p for trend was examined between week 1 (initial device placement visit) and week 5 (1 week after PENFS completion) as well as between week 1 and 3-month follow-up.

In addition, associations between questionnaire scores and anthropometric measures were examined longitudinally with an interaction between each device placement visit and questionnaire score. Weight and BMI were examined in a log scale and all results were presented as β- coefficients with 95% CI. All analyses were conducted as two-sided tests using p≤0.05 as statistically significant using SAS version 9.4 (SAS Institute Inc., Cary, North Carolina).

3 |. Results

A total of 116 patients with FD were identified that completed PE NFS treatment (Table 1), and 90/116 (78%) reported restricted oral intake (mean age 16.1 ± 2.6 years and median age 16years [15–18]). Eighty-seven percent of patients with FD and restricted oral intake were female and 91% were Caucasian. Of the cohort, 29% had epigastric pain FD-subtype, 20% had postprandial distress FD-subtype, and 65% had both. Most common comorbidities were anxiety (91%), orthostatic intolerance (77%), depression (71%), sleep disturbance (63%), and joint hypermobility syndrome (52%). Thirty-seven percent of patients with FD and restricted oral intake were on medications with known weight-altering side effects. However, they remained on the same stable dosing during the standard PENFS treatment protocol.

TABLE 1 |.

Demographic characteristics of functional dyspepsia patients with restricted oral intake.

Demographics		Restricted oral intake (n = 90)
	Age [Mean (y) ± SD]	16.1 ± 2.60
Gender [n (%)]	Female	78 (86.67)
	Male	12 (13.33)
Race [n (%)]	White	82 (91.11)
	Black	2 (2.22)
	Asian or Middle eastern	1 (1.11)
	Mixed	5 (5.56)
FD subtype [n (%)]	Epigastric pain syndrome	26 (28.88)
	Postprandial distress syndrome	18 (20.00)
	Both	46 (51.11)
Additional DGBI [n (%)]	Irritable bowel syndrome	44 (48.89)
	Rumination	7 (7.78)
	Functional abdominal pain-not otherwise specified	6 (6.67)
	Functional nausea/vomiting	4 (4.44)
Common comorbidities	Anxiety	82 (91.11)
	Orthostatic intolerance3	69 (76.66)
	Depression	64 (71.11)
	Sleep disturbance	57 (63.33)
	Hypermobility spectrum disorder	47 (52.22)
	Post-traumatic stress disorder (PTSD)	14 (15.56)
	Bipolar disorder	7 (7.78)

^aOrthostatic intolerance includes diagnoses of orthostatic hypotension, non-cardiac dizziness, vasovagal syncope, dysautonomia, and Postural orthostatic tachycardia syndrome.

At baseline, 42% had recorded weight loss (lower weight at initial device placement visit compared to last gastroenterology clinic visit) prior to start of PENFS and 13% required enteral nutrition support (via nasogastric, nasojejunal, gastrostomy, jejunostomy or gastro-jejunostomy feeds). Restricted eating was secondary to nausea (94%), pain (93%), vomiting (26%), and bloating (16%). Fifty-one percent restricted the amount of food intake, and 41% endorsed decreased appetite (Table 2).

TABLE 2 |.

Restricted eating habits of functional dyspepsia patients.

Restricted eating behaviors (n = 90)
Decreased amount of oral intake	46 (51%)
Decreased appetite	37 (41%)
Picky eating/limited preferred foods	6 (7%)
Pushing oneself to eat	5 (6%)
Restricted eating due to food sensory characteristics	3 (3%)
Restricted eating due to generalized gastrointestinal symptoms	72 (80%)
Restricted eating due to nausea	85 (94%)
Restricted eating due to pain	84 (93%)
Restricted eating due to vomiting	23 (26%)
Restricted eating due to bloating	14 (16%)

API, NSS, PCS-C, FDI, PHQ-9, and CSI scores (total, GI, and non-GI sub-scales) longitudinally improved after PENFS completion and at 3 months (p < 0.0001 for all, Figure 1). Weight and BMI also showed improvement from baseline to post-treatment (p = 0.02 and 0.002 respectively) as well as from baseline to 3-month follow-up (p =0.004 and p = 0.003 respectively). BMI Z-scores trended to show an improvement from baseline to treatment completion (p =0.07) and significantly improved from baseline to 3-month follow-up (p = 0.03, Figure 2).

FIGURE 1 |.

Changes in outcomes in functional dyspepsia patients with restricted eating. Examined via mixed linear regression models with placement visits as fixed effect and subject modeled as random effect to account for repeated measure. Results for group differences presented as Least Square Means (LS Means) and 95%confidence interval. API, Abdominal Pain Index; CSI, Children’s Somatization Inventory; FDI, Functional Disability Inventory; Non-GI & GI, Non-gastrointestinal and gastrointestinal; NSS, Nausea Severity Scale; PCQ, Pain Coping Questionnaire; PCS-C, Pain catastrophizing; PENFS, Percutaneous Electrical Nerve Field Stimulation; PHQ-9, Patient Health Questionnaire; PISI, Pediatric Insomnia Severity Index; PROMIS-Anxiety, Patient Reported Outcomes Measurement Systems-Anxiety.

FIGURE 2 |.

Changes in anthropometric outcomes in functional dyspepsia patients with restricted eating. Mixed linear regression models used. Results presented as Least Square Means (LS Means) and 95%confidence interval (CI). BMI, body-mass-index. Weight Z-score and BMI Z-score are per World Health Organization (WHO) calculations. 3Month FU—3month follow up post-PENFS treatment.

A significantly negative association was found between NSS, PCS-C, PHQ-9, CSI, and CSI-GI and weight-for-age Z-scores during weeks 2–4 in comparison to the initial device placement visit (Table 3, p < 0.05 for all). No significant associations were found between clinical outcomes and weight, BMI, and BMI Z-scores (Supplemental Table S1).

TABLE 3 |.

Longitudinal association between clinical and anthropometric outcomes of functional dyspepsia patients with restricted oral intake undergoing PENFS.

Independent variables		Weight-for-age Z-scoresa		Body mass index Z-scoresa
		β-coefficient (LCL, UCL)	p value	β-coefficient (LCL, UCL)	p value
API	Placement Visit		0.0764		0.8355
	Week 2	0.112 (−0.151, 0.375)		0.035 (−0.081, 0.152)
	Week 3	0.307 (0.042, 0.572)		−0.007 (−0.124, 0.111)
	Week 4	0.266 (−0.001, 0.534)		0.003 (−0.114, 0.120)
	API Score (Week 1)	118.780 (12.428, 225.130)	0.1022	22.395 (−24.565, 69.355)	0.2542
	API × Placement Visit		0.3267		0.3209
	API at Week 2	−44.990 (−148.000, 58.017)		−24.935 (−70.954, 21.085)
	API at Week 3	−93.885 (−199.950, 12.183)		5.000 (−42.683, 52.683)
	API at Week 4	−77.582 (−184.960, 29.792)		9.989 (−37.709, 57.686)
NSS	Placement Visit		0.0003 *		0.618
	Week 2	0.030 (−0.202, 0.262)		−0.014 (−0.122, 0.095)
	Week 3	0.369 (0.134, 0.603)		0.036 (−0.073, 0.145)
	Week 4	0.339 (0.101, 0.576)		0.036 (−0.075, 0.147)
	NSS Score (Week 1)	76.082 (−26.192, 178.360)	0.9252	16.531 (−33.807, 66.869)	0.6603
	NSS × Placement Visit		0.0017 *		0.8428
	NSS at Week 2	−10.183 (−107.270, 86.905)		−4.246 (−49.978, 41.485)
	NSS at Week 3	−141.640 (−239.020, −44.262)		−18.763 (−64.784, 27.258)
	NSS at Week 4	−137.230 (−240.070, −34.388)		−9.174 (−57.305, 38.958)
PCS-C	Placement Visit		0.0014 *		0.3464
	Week 2	0.114 (−0.136, 0.363)		−0.014 (−0.119, 0.092)
	Week 3	0.416 (0.151, 0.682)		−0.074 (−0.185, 0.037)
	Week 4	0.407 (0.147, 0.668)		−0.005 (−0.115, 0.105)
	PCS Score (Week 1)	13.664 (0.453, 26.874)	0.4086	−1.023 (−6.699, 4.654)	0.8954
	PCS × Placement Visit		0.0072 *		0.3124
	PCS at Week 2	−5.004 (−16.163, 6.155)		−0.655 (−5.403, 4.094)
	PCS at Week 3	−16.143 (−27.903, −4.383)		3.126 (−1.870, 8.122)
	PCS at Week 4	−16.945 (−28.713, −5.177)		0.476 (−4.500, 5.452)
FDI	Placement Visit		0.0734		0.9473
	Week 2	−0.004 (−0.150, 0.141)		0.004 (−0.068, 0.075)
	Week 3	0.134 (−0.007, 0.274)		−0.014 (−0.084, 0.055)
	Week 4	0.117 (−0.023, 0.258)		−0.005 (−0.074, 0.064)
	FDI Score (Week 1)	4.502 (−2.524, 11.528)	0.4762	−0.744 (−4.354, 2.866)	0.4367
	FDI × Placement Visit		0.1718		0.8939
	FDI at Week 2	0.156 (−6.068, 6.380)		−1.139 (−4.219, 1.941)
	FDI at Week 3	−4.929 (−10.884, 1.027)		−0.357 (−3.330, 2.616)
	FDI at Week 4	−4.508 (−10.578, 1.562)		−0.313 (−3.352, 2.727)
PISI	Placement Visit		0.0462 *		0.4103
	Week 2	−0.069 (−0.329, 0.192)		−0.055 (−0.167, 0.058)
	Week 3	0.241 (−0.022, 0.505)		−0.071 (−0.184, 0.042)
	Week 4	0.238 (−0.035, 0.510)		0.011 (−0.105, 0.128)
	PISI Score (Week 1)	6.499 (−13.038, 26.036)	0.8769	7.018 (−1.492, 15.528)	0.0297
	PISI × Placement Visit		0.1913		0.6604
	PISI at Week 2	3.304 (−12.811, 19.420)		2.036 (−5.039, 9.110)
	PISI at Week 3	−12.038 (−28.521, 4.445)		4.582 (−2.667, 11.831)
	PISI at Week 4	−11.409 (−28.463, 5.645)		1.887 (−5.612, 9.385)
SCARED	Placement Visit		0.2896		0.5179
	Week 2	−0.032 (−0.081, 0.018)		−0.016 (−0.077, 0.044)
	Week 3	−0.047 (−0.097, 0.004)		−0.026 (−0.089, 0.036)
	Week 4	−0.043 (−0.099, 0.012)		−0.050 (−0.119, 0.018)
	SCARED Score (Week 1)	−1.462 (−3.506, 0.583)	0.1527	−2.536 (−5.037, −0.034)	0.0274
	SCARED × Placement Visit		0.6809		0.7241
	SCARED at Week 2	0.895 (−0.921, 2.712)		0.543 (−1.686, 2.771)
	SCARED at Week 3	0.202 (−1.391, 1.795)		−0.619 (−2.574, 1.337)
	SCARED at Week 4	−0.300 (−2.103, 1.503)		0.500 (−1.712, 2.711)
PHQ-9	Placement Visit		0.0303 *		0.631
	Week 2	0.080 (−0.137, 0.297)		−0.030 (−0.125, 0.066)
	Week 3	0.306 (0.079, 0.534)		−0.008 (−0.104, 0.088)
	Week 4	0.261 (0.032, 0.491)		0.031 (−0.070, 0.132)
	PHQ9 Score (Week 1)	−4.639 (−24.832, 15.554)	0.0111	3.185 (−5.933, 12.304)	0.4118
	PHQ9 × Placement Visit		0.0407 *		0.9134
	PHQ9 at Week 2	−11.523 (−33.632, 10.587)		1.424 (−8.207, 11.054)
	PHQ9 at Week 3	−30.622 (−54.011, −7.233)		0.510 (−9.203, 10.223)
	PHQ9 at Week 4	−27.801 (−51.402, −4.200)		−2.115 (−12.263, 8.034)
PCQ	Placement Visit		0.1927		0.3963
	Week 2	0.165 (−0.375, 0.705)		0.168 (−0.772, 1.108)
	Week 3	−0.145 (−0.631, 0.341)		−0.225 (−1.070, 0.620)
	Week 4	−0.122 (−0.633, 0.389)		−0.205 (−1.091, 0.682)
	PCQ Score (Week 1)	10.429 (−44.346, 65.204)	0.2568	−1.913 (−97.215, 93.389)	0.9107
	PCQ × Placement Visit		0.2322		0.5427
	PCQ at Week 2	−20.881 (−87.989, 46.226)		−20.173 (−136.970, 96.620)
	PCQ at Week 3	14.165 (−47.925, 76.254)		17.656 (−89.850, 125.160)
	PCQ at Week 4	10.258 (−53.202, 73.719)		17.575 (−92.346, 127.500)
PROMIS-Anxiety	Placement Visit		0.7367		0.9728
	Week 2	−0.122 (−0.353, 0.109)		0.020 (−0.227, 0.267)
	Week 3	−0.093 (−0.345, 0.159)		0.059 (−0.211, 0.328)
	Week 4	−0.063 (−0.320, 0.194)		0.037 (−0.237, 0.312)
	PROMIS-Anxiety Score (Week 1)	−1.889 (−7.997, 4.220)	0.7169	−0.380 (−6.907, 6.147)	0.6198
	PROMIS-Anxiety × Placement Visit		0.7779		0.9055
	PROMIS-Anxiety at Week 2	2.294 (−2.313, 6.900)		−0.317 (−5.249, 4.615)
	PROMIS-Anxiety at Week 3	1.250 (−3.702, 6.202)		−1.735 (−7.036, 3.566)
	PROMIS-Anxiety at Week 4	0.862 (−3.993, 5.718)		−1.038 (−6.233, 4.158)
CSI	Placement Visit		0.011 *		0.6505
	Week 2	0.002 (−0.227, 0.232)		−0.050 (−0.154, 0.054)
	Week 3	0.290 (0.060, 0.520)		−0.002 (−0.105, 0.102)
	Week 4	0.237 (0.000, 0.474)		−0.003 (−0.110, 0.105)
	CSI Score (Week 1)	4.301 (−4.064, 12.665)	0.7686	−0.305 (−4.165, 3.555)	0.9568
	CSI × Placement Visit		0.0363 *		0.8788
	CSI at Week 2	0.077 (−6.481, 6.634)		0.734 (−2.217, 3.686)
	CSI at Week 3	−7.363 (−13.767, −0.959)		−0.284 (−3.158, 2.589)
	CSI at Week 4	−5.956 (−12.502, 0.589)		0.431 (−2.551, 3.413)
CSI-GI	Placement Visit		0.0085 *		0.9301
	Week 2	−0.022 (−0.313, 0.269)		−0.036 (−0.170, 0.097)
	Week 3	0.324 (0.032, 0.615)		−0.009 (−0.142, 0.124)
	Week 4	0.215 (−0.080, 0.511)		−0.019 (−0.153, 0.116)
	CSI-GI Score (Week 1)	1.777 (−19.850, 23.403)	0.2989	−2.339 (−12.397, 7.719)	0.5805
	CSI-GI × Placement Visit		0.0232 *		0.9366
	CSI-GI at Week 2	0.948 (−19.213, 21.110)		0.407 (−8.917, 9.732)
	CSI-GI at Week 3	−22.046 (−42.233, −1.859)		−0.298 (−9.606, 9.009)
	CSI-GI at Week 4	−15.327 (−36.170, 5.516)		2.049 (−7.535, 11.632)
CSI-Non GI	Placement visit		0.0264 *		0.4284
	Week 2	−0.007 (−0.180, 0.166)		−0.046 (−0.124, 0.032)
	Week 3	0.197 (0.024, 0.370)		0.001 (−0.076, 0.079)
	Week 4	0.174 (−0.002, 0.351)		0.009 (−0.072, 0.089)
	CSI-non GI Score (Week 1)	6.779 (−3.745, 17.303)	0.4417	0.444 (−4.430, 5.317)	0.7453
	CSI-non GI × Placement Visit		0.1366		0.8232
	CSI-non GI at Week 2	0.390 (−7.575, 8.356)		1.192 (−2.359, 4.742)
	CSI-non GI at Week 3	−6.888 (−14.586, 0.810)		−0.414 (−3.856, 3.028)
	CSI-non GI at Week 4	−6.311 (−14.179, 1.557)	.	0.257 (−3.287, 3.802)	.

Abbreviations: API, Abdominal Pain Index; CSI, Children’s Somatization Inventory; FDI, Functional Disability Inventory; GI & Non-GI, Gastrointestinal & Non-gastrointestinal; NSS, Nausea Severity Scale; PCQ, Pain Coping Questionnaire; PCS-C, Pain catastrophizing; PISI, Pediatric Insomnia Severity Index PHQ-9: Patient Health Questionnaire; PROM IS-Anxiety, Patient Reported Outcomes Measurement Systems-Anxiety.

^*p ≤ 0.05.

^aExamined in log scale.

Compared to baseline, patients with FD and restricted oral intake reported significantly increased food intake at the final visit (37% vs. 13%, p =0.01). Of the patients relaying solely on enteral tube feedings, 5/11 (45%) had improved oral intake after PENFS completion (Table 4). Seventy-three percent had improved the variety of oral food intake while 2 patients (Table 4, Patient D and H) were able to completely wean off of enteral nutrition. Two patients (Table 4, Patient B and G) had major enteral nutrition changes with a significant decrease in enteral nutrition use or consolidation of feeds with higher volume tolerance.

TABLE 4 |.

Changes in tube feedings and oral intake with functional dyspepsia patients with restricted eating and enteral nutrition dependence after PENFS.

Patient	Age (y)	Gender	Mode of enteral nutrition	Rate	Type of formula	Baseline % calories through enteral feedings	Baseline oral intake (yes/no)	Improved amount of oral intake	Improved variety of oral intake	New foods added	Post PENFS rate of enteral nutrition
A	23	Female	Nasogastric	0	Ensure Clear	97%	No	Major	Yes	Not specified	0
B	24	Female	Nasojejunal	95 mL/h × 8.5 h	Boost Breeze	68%	Yes	Slight	Yes	Sandwich, chips, lasagna	95 mL/h × 1.5 h
C	17	Female	Nasogastric	45 mL/h × 8 h	Kate Farms 1.0	18%	Yes	Same	No	None	45 mL/h × 16 h
D	18	Female	Nasogastric	90 mL/h × 8 h	Boost Plus	100% (not including snacks)	No	Major	Yes	Pancakes, yogurt, fish, sandwich, pizza, chicken nuggets	0
E	14	Female	Nasojejunal	90 mL/h × 6 h	Peptamen Jr	50%	Yes	Slight	Yes	Not specified	Not available (no follow-up)
F	22	Male	Gastrojejunal	90 mL/h × 16 h	Jevity 1.5	100% (not including snacks)	No	None	Yes	None	0
G	19	Female	Gastrojejunal, intravenous fluids	50 mL/h × 22 h	50% Kate Farms Peptide 1.5 kcal/oz	50%	Yes	Slight	Yes	Bread, sausage, ham, thicker soups	100 mL/h × 10 h
H	20	Female	Nasogastric	225 mL/h × 2 h	Osmolite 1.5 + Pedialyte + water	100% (not including snacks)	Yes	Major	Yes	Eggs, dessert	0
I	18	Female	Nasojejunal	0	0	0	No	Major	Yes	Not specified	0
J	18	Female	Gastrojejunal	83 mL/h × 20 h	Pediasure 1.0	100%	No	None	No	None	Not available (no follow-up)
K	18	Female	Gastrostomy tube	96 mL/h × 20 h	Jevity	100%	No	Slight	Yes	Mashed potatoes, ice cream, soup	Not available (no follow-up)

4 |. Discussion

This study is the first to systematically assess anthropometric and nutritional outcomes following completion of PENFS and investigate the efficacy of neurostimulation in patients with functional dyspepsia requiring enteral nutrition support. We found that PENFS significantly improved symptoms, anthropometrics, and oral intake in patients with FD who restricted their oral intake.

Our study identified 78% of children and adolescents with FD to have restricted oral intake. Our study has similar rates of EPS (29%) compared to global prevalence (28%) but lower PDS (20%) and higher overlap (65%) compared to global prevalence (67% and 12% respectively) [20]. Untreated ARFID can ultimately lead to malnutrition and stunted growth as well as poor quality of life and prolonged comorbid anxiety and depression [21]. Restrictive eating rates in our study are similar to a recent study by Kaul et al. where at least 65% of pediatric patients with FD and 49% with gastroparesis endorsed ARFID [22]. Previous work from our group has shown that 50% of children with DGBI had restrictive oral intake. Those with restricted intake had higher rates and intensity of nausea and depression, significantly increased prevalence of comorbid conditions such as attention-deficit/hyperactivity disorder and sleep disturbances, as well as lower baseline body weight compared to those without restrictive eating [23]. Similarly, Murray et al. showed that in a subset of older patients with DGBI and ARFID, older age and lower weight status via BMI were significantly associated with ARFID [24]. Weeks et al. demonstrated that 13%–40% of DGBI patients meet partial to clinically significant ARFID criteria [25]. Murray et al. also reviewed a cohort of 410 adult patients with neurogastroenterology and motility disorders [26]. They found that 6% of patients met full ARFID criteria and 17% had clinically significant avoidant or restrictive eating behaviors [26]. Such prevalence rates are lower than pediatric DGBI cohorts, highlighting the need for treatment interventions in pediatric FD and co-occurring ARFID.

Cognitive behavioral therapy and standard medication therapy (e.g., tricyclic antidepressant, serotonin receptor antagonist) are often offered to resolve mild symptoms, but often not when complicated by nutritional issues such as weight loss [27]. Neuromodulation has become a rapidly evolving option for DGBI, including FD. Use of PENFS has shown both improvements in gastrointestinal symptoms such as abdominal pain and nausea, pain catastrophizing, and somatization as well as psychosocial habits such as depression, anxiety and functional disability [28].

PENFS has shown improvements in multiple domains of patients with FD. Compared to standard medical therapy, PENFS demonstrated improvement in abdominal pain, nausea, disability and anxiety/depression that was sustained for 3 months [9]. According to a multicenter registry, significant improvements were noted in abdominal pain and nausea at 3 weeks and 3 months with PENFS therapy with sustained effects up to 12 months in a subset (29]. Following 4 weeks of PENFS therapy, there was a decrease in anxiety and functional impairment (29].

In our study, we were also able to affirm previously reported symptom improvement after PENFS treatment, including abdominal pain, nausea, pain catastrophizing, functional disability, and somatization, and demonstrate an association for the first time with anthropometric changes [9]. Weaning off of enteral nutrition and improved oral intake for solely enteral tube fed patients is a remarkable breakthrough for the treatment of such complex DGBI associated with feeding intolerance.

Our study noted a negative association between NSS, PCS-C, PHQ-9, CSI, and CSI-GI and weight-for-age Z-scores. We postulated certain mechanisms by which PENFS showed improvements in restrictive eating patterns. It is very likely that patients improved their oral intake due to improvement in symptoms with PENFS therapy. However, it is also possible that PENFS might have a direct effect on the hypothalamic satiety centers. Its mechanism of action on the amygdala has been shown in animal models and, as such, it could have further effects on related limbic system structures [8]. This is supported by the fact that eating habits, oral intake, and their relationship to the brain have been greatly studied, as the vagal nerve critically communicates bidirectionally between the gut and hepatobiliary system to the CNS through information provided by chemo-mechanical receptors that detect food and appetite and satiety hormones ghrelin and cholecystokinin [30–32]. Similarly, non-invasive neuromodulation, such as transcranial magnetic stimulation (TMS), has been used in occasional severe, difficult-to-treat eating disorders (ED) such as anorexia nervosa and binge eating disorder in order to target craving levels effectively (33, 34]. It is also theorized that brain activity and response to neuromodulation can vary in relation to one’s metabolic state, which can be affected in ED patients with poor nutrition (35].

An additional possible mechanism to consider involves the microbiota-gut-brain axis, in which the vagal nerve both senses microbiota metabolite changes via afferent branches and transfers such information to the CNS via efferent signaling which in causes a response, including the notable cholinergic anti-inflammatory pathway that can decrease intestinal permeability. Stressors onto the vagal nerve that cause poor vagal tone can ultimately cause peripheral inflammation; therefore, VNS arguably contains anti-inflammatory properties that can prevent dysbiosis (36]. Furthermore, CNS-based depression and stressors can induce gut hyperpermeability and inflammation, and the feedback of low vagal tone can cause worsened mental health, suggesting a bidirectionality relationship [37]. Animal models have demonstrated this point, where one study on healthy rats with inoculated fecal samples from rats experiencing chronic mild stress demonstrated early sustained neuroinflammatory changes in hippocampal neurogenesis related to depression (38]. Healthy rats that underwent a vagal nerve dissection ultimately experienced no CNS changes, suggestive of the vagal afferent nerve integrity being critical to mood changes. Thus, improvements in mood and sleep with use of PENFS could also lead to improved nutritional intake.

Similar to our study, brief bursts of vagal nerve stimulation (VNS) through transcranial direct current stimulation (tDCS) have resulted in mild to moderate improvement of oropharyngeal muscle use and swallow function in stroke patients and oropharyngeal dysphagia populations. This could be secondary to synaptic reorganization of orofacial motor networks (39, 40]. This opens a new line of investigation and applicability of non-invasive neuromodulation for treating eating disorders and disordered eating. Future clinical trials can focus on specific treatment outcomes geared towards this population.

Advantages to our study include a relatively large sample size that has been reliably followed within the same institution and utilization of data that has been documented by a clinically consistent provider group protocol. A demonstration of improved gastrointestinal symptoms and decreased restricted eating habits after 3 months is also beneficial in examining both short- and long-term outcomes in patients with FD undergoing PENFS.

Limitations to consider include the retrospective design, absence of a matching cohort of patients with FD without restricted oral intake, as such patients may represent a milder symptom burden that may not demonstrate as robust results as our cohort. Alternatively, the study did not isolate findings to patients with pre-diagnosed ARFID as such restrictions would have likely decreased the sample size and missed a large portion of patients who did not carry a specific diagnosis but have symptoms of restricted intake. Prospective studies and clinical trials may provide data on longitudinal outcomes. Pre-clinical studies could aid in discovering mechanisms and enhance current treatment approaches. Furthermore, patients with FD and restricted oral intake have other comorbid conditions (i.e., hyper mobility spectrum disorder, orthostatic intolerance, chronic pain conditions) that may contribute to overall clinical outcomes to a significant degree. As a result, it is critical to consider the stratification of such patients and their specific responses to treatment as it may reveal the benefit or threshold to which PENFS can be useful for patients DGBI and other outcomes. Future studies should consider the manner in which oral intake is monitored, whether it is based in anthropometric measures, self-reported diaries, or close dietician monitoring.

In conclusion, PENFS should be considered as a viable treatment for patients with FD that often develop or already present restrictive eating. Robust clinical trials of neuromodulation could serve to improve and optimize neuromodulation strategies for disordered eating.

Supplementary Material

Supporting Information

Additional supporting information can be found online in the Supporting Information section. Table S1: Longitudinal association between clinical scores and anthropometric outcomes of functional dyspepsia patients with restricted oral intake undergoing PENFS.

Key Points.

• Pediatric patients with functional dyspepsia (FD) often restrict oral intake due to gastrointestinal symptoms and lose weight.

• PENFS can improve gastrointestinal symptoms and psychosocial outcomes in children with disorders of gut-brain interaction.

• PENFS improved weight, BMI, and oral intake with decreased dependence on enteral nutrition feeds in addition to improving clinical symptoms in patients with FD, including children, and restricted oral intake.

Abbreviations

API

Abdominal Pain Index

ARFID

avoidant/restrictive food intake disorder

BMI

Body Mass Index

CNS

central nervous system

CSI

Children’s Somatization Inventory

DGBI

disorders of gut-brain interaction

EPS

epigastric pain syndrome

FD

functional dyspepsia

FDI

Functional Disability Inventory

GI

gastrointestinal

LS Means

Least Square Means

NIAS

Nine Item Avoidant/Restrictive Food Intake Disorder (ARFID) Screen

NSS

Nausea Severity Scale

PCQ

Pain Coping Questionnaire

PCS-C

Pain catastrophizing

PDS

postprandial distress

PENFS

peripheral electrical nerve field stimulation

PHQ-9

Depression

PISI

Pediatric Insomnia Severity Index

POTS

postural orthostatic tachycardia syndrome

PROMIS-Anxiety

Anxiety-Patient Reported Outcomes Measurement Systems

tDCS

Transcranial Direct Current Stimulation

TMS

transcranial magnetic stimulation

VNS

vagal neurostimulation

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.